Why Employers Make Bad Choice Architects

Federal retirement policy has long been premised on the view that many of us, if left to our own devices, will save too little for retirement. A growing literature in behavioral economics has shown that seemingly small nudges in employer retirement plan design, like automatically enrolling workers into contributing to the plan, can have large effects on behavior. Many have seized on these findings to advocate that employers design the “choice architecture” of their 401(k) plans in order to improve their workers’ choices.

Indeed, this approach is widely heralded as the most successful application of behavioral economics to public policy to date. The Pension Protection Act of 2006 created regulatory incentives for employers to adopt automatic enrollment, and employers have done so in droves.

In a recent article titled A Behavioral Contract Theory Perspective on Retirement Savings, we argue that casting the employer in the role of choice architect is misguided. Employers offer retirement plans to attract workers in the labor market. The operation of the labor market creates powerful incentives for plan design that depend on the intertemporal preferences and biases of workers. We show that if those workers make systematic mistakes in their retirement savings decisions, then the labor market will produce incentives for plan designs that generally fail to effectively address the problems. Indeed, the presence of workers who undersave due to myopia results in employer plan designs that exploit the myopic by lowering their compensation. Our analytic framework provides novel explanations for a range of features of plan design, including the high prevalence of matching contributions, the use of low default contribution rates in automatic enrollment plans, the shift away from annuities toward lump sum distributions, and the offering of investment options with excessive fees. The regulation of these plans should be reformed to address the problems we identify. More fundamentally, our analysis calls for a rethinking of the current scheme’s special subsidies for employer-sponsored plans.

To illustrate our analysis, consider two key facts about 401(k) plan design. First, the vast majority of plans (over 80%) offer employer matching contributions, in which for every dollar the employee contributes, the employer promises to kick in a dollar, say, typically up to some cap like 5% of salary. Second, at employers that offer matching, many employees (on the order of 40%) fail to contribute enough to receive the maximum match.

We argue that both of these facts can be explained using the standard explanation in behavioral economics for the undersaving problem, which is that workers are myopic or present-biased. In the labor market, naïve myopic workers overvalue employers’ offers to make matching contributions because they overestimate how much they will save under the match. Offering matching contributions thus gives employers more bang for their buck in attracting workers than offering a higher salary. One result is that naïve myopic workers cross-subsidize rational workers who in fact will save enough to receive the full match. The cross-subsidy created by matching results in naïve myopic workers being paid less than they would otherwise. So matching in a sense exploits naïve myopic workers: it results in them receiving less total compensation. The irony here is that these are the very type of workers that federal retirement savings policy—including the special subsidies for employer-sponsored retirement plans—seeks to help.

Moreover, the seeming success of employer adoption of automatic enrollment and Save More Tomorrow-type schemes is largely illusory. The adoption of automatic enrollment by employers has in fact reduced the retirement savings of many households and may in fact have reduced overall retirement savings. The reason is that any automatic enrollment plan must, by definition, have a default contribution rate. The most common default contribution rate chosen by employers is three percent of salary. So the positive effect that automatic enrollment has on participation rates can be thought of as moving some workers who would have initially contributed zero percent to the plan to contributing three percent to the plan. But the default contribution rate is also sticky for workers who would have enrolled on their own in a traditional opt-in design, including those who would have enrolled at rates far higher than three percent. For these households, automatic enrollment has reduced retirement savings. On net it appears that this latter negative effect on savings outweighs the positive effect, as reflected in the fact that average savings rates in 401(k) plans have fallen over the same period that employers adopted automatic enrollment.

Why have employers adopted these low default contribution rates in automatic enrollment plans?   Our analytic framework provides an explanation. Because workers are insensitive to the default contribution rate when choosing among competing compensation contracts in the labor market—this level of detail in plan design is hardly salient to the typical worker—employers have strong incentives to choose the default contribution rate that minimizes contributions to the plan, since with the resulting lower employer matching contributions, the employer can then offer a higher wage.   You might intuitively think that that default contribution rate is the smallest one possible, i.e., zero, as in a traditional opt-in plan. But it is in fact a (low) non-zero contribution rate, since such a default reduces the savings of workers who, under an opt-in design, would enroll at a higher contribution rate, as has been empirically documented in the literature. Our behavioral contract theory thus explains the pattern of low default contribution rates combined with matching that employers in fact commonly choose.

The upshot of our analysis is that the operation of the labor market produces bad incentives for employers as choice architects for retirement savings. The very behavioral biases that produce the undersaving problem that federal retirement savings policy is intended to address also undermine the functioning of employers as choice architects. Those biases induce preferences of workers across alternative plan designs that workers express through their choices in the labor market. The operation of the labor market then leads employers to choose plan designs that not only do not effectively address the undersaving problem, they exacerbate it by reducing the compensation of the very workers prone to undersaving.

One potential policy response to the problems with employer incentives that we identify would be to reform the regulation of employer-sponsored plans. For example, currently these plans are subject to a set of nondiscrimination rules that actively encourage employers to offer matching. At a minimum, we think these rules should be reformed to not encourage matching. But perhaps a better approach would be to simply ban matching and require that all employer contributions take a non-elective form.

More fundamentally, our analysis calls into question the longstanding delegation of choice architecture in this area to employers. Employers play this role largely as a consequence of the preferential tax treatment of employer-sponsored retirement plans under current law. A central insight of our article is that employers have incentives that conflict with the goal of providing myopic workers with an optimal retirement savings choice architecture. This suggests a more far-reaching overhaul of retirement savings policy: supplanting employers as choice architects. We could do so by scrapping the tax subisidies for employer-sponsored plans and creating in their place a federally-sponsored defined contribution savings vehicle, supplemental to Social Security, available to all workers independent of their employer.

The preceding post comes to us from Ryan Bubb, Professor of Law at New York University School of Law, Patrick Corrigan, J.D. expected from New York University School of Law, and Patrick L. Warren, Associate Professor of Economics at Clemson University.  The post is based on their recent article, which is entitled “A Behavioral Contract Theory Perspective on Retirement Savings” and available here.